A new bill introduced in federal parliament by independent Senator David Pocock seeks to criminalise the sharing of non-consensual AI-generated deepfakes in Australia, amidst growing concern over the misuse of artificial intelligence and the government's perceived inaction. The legislation would significantly expand protections for individuals whose likeness—such as face, voice, or mannerisms—is replicated by AI without consent, by creating legal recourse and instituting strict penalties.
The bill proposes key amendments to the Online Safety Act, giving the eSafety Commissioner expanded powers to demand the removal of non-consensual deepfakes and issue fines. Individuals could be fined $165,000, while companies failing to act on removal notices might face penalties up to $825,000. Repeat offenders who ignore takedown requests could also receive additional fines. In parallel, changes to the Privacy Act would enable victims to bring civil lawsuits against perpetrators without having to prove financial loss—recognising the emotional and reputational harm such deepfakes can cause.
Senator Pocock framed the bill as a necessary response to the rapid rise of generative AI and the current legal gap that leaves Australians largely unprotected unless the content is sexually explicit. “This bill would put into legislation that your face, voice and likeness is yours,” he said. He added that current safeguards are insufficient, leaving people vulnerable to disinformation, scams, and commercial exploitation. “To me, this seems like a bit of a no-brainer. You should own your face.”
Importantly, the bill includes exemptions for journalism, satire, law enforcement, and children, and is designed to protect those who unknowingly share deepfakes while targeting deliberate or reckless dissemination. Pocock noted the bill also applies to political campaigns, pointing to his own AI-generated videos of political leaders used during the last election to raise awareness about the risks of AI manipulation. “We have to draw a line in the sand and say, this is not on — you cannot deepfake someone without their consent,” he stated.
Independent MP Kate Chaney has also taken legislative steps to address AI-related harms. In July, she introduced a private member’s bill aimed at criminalising the use of AI tools designed to create child sexual abuse material. Despite acknowledgement from the attorney-general’s office that this is a concern, Chaney criticised the lack of governmental follow-through. “Still, we’ve seen no action on it,” she said.
Both Chaney and Pocock have been vocal about the need for comprehensive AI regulation, with Chaney calling for the establishment of an Australian AI Safety Institute. “The US, the UK, Canada, Japan, Singapore all have an equivalent. Australia has supported the idea of them but has not yet actually taken action,” she warned.
While the federal government passed a law in August 2024 targeting non-consensual sexually explicit deepfakes, broader protections remain absent. The current administration has shifted focus towards the economic benefits of AI, with Minister Tim Ayres stating the government would wait for a comprehensive review due in late 2025 before taking legislative steps. Pocock and Chaney, however, argue that waiting is no longer an option. “This is a huge freight train that is coming at us and we’re unprepared,” Pocock said.